Original Research published: 24 July 2017 doi: 10.3389/fimmu.2017.00823
Free Fatty acids Profiles are related to gut Microbiota signatures and short-chain Fatty acids Javier Rodríguez-Carrio1, Nuria Salazar1, Abelardo Margolles1, Sonia González2, Miguel Gueimonde1, Clara G. de los Reyes-Gavilán1* and Ana Suárez3 Department of Microbiology and Biochemistry of Dairy Products, Instituto de Productos Lácteos de Asturias (IPLA-CSIC), Villaviciosa, Asturias, Spain, 2 Area of Physiology, Department of Functional Biology, University of Oviedo, Oviedo, Asturias, Spain, 3 Area of Immunology, Department of Functional Biology, University of Oviedo, Oviedo, Asturias, Spain 1
Edited by: Marina I. Arleevskaya, Kazan State Medical Academy, Russia Reviewed by: Mario M. D’Elios, University of Florence, Italy Francois-Pierre Martin, Nestle Institute of Health Sciences, Switzerland Arne Yndestad, Oslo University Hospital, Norway *Correspondence: Clara G. de los Reyes-Gavilán [email protected] Specialty section: This article was submitted to Microbial Immunology, a section of the journal Frontiers in Immunology Received: 07 April 2017 Accepted: 29 June 2017 Published: 24 July 2017 Citation: Rodríguez-Carrio J, Salazar N, Margolles A, González S, Gueimonde M, de los ReyesGavilán CG and Suárez A (2017) Free Fatty Acids Profiles Are Related to Gut Microbiota Signatures and Short-Chain Fatty Acids. Front. Immunol. 8:823. doi: 10.3389/fimmu.2017.00823
A growing body of evidence highlights the relevance of free fatty acids (FFA) for human health, and their role in the cross talk between the metabolic status and immune system. Altered serum FFA profiles are related to several metabolic conditions, although the underlying mechanisms remain unclear. Recent studies have highlighted the link between gut microbiota and host metabolism. However, although most of the studies have focused on different clinical conditions, evidence on the role of these mediators in healthy populations is lacking. Therefore, we have addressed the analysis of the relationship among gut microbial populations, short-chain fatty acid (SCFA) production, FFA levels, and immune mediators (IFNγ, IL-6, and MCP-1) in 101 human adults from the general Spanish population. Levels of selected microbial groups, representing the major phylogenetic types present in the human intestinal microbiota, were determined by quantitative PCR. Our results showed that the intestinal abundance of Akkermansia was the main predictor of total FFA serum levels, displaying a negative association with total FFA and the pro-inflammatory cytokine IL-6. Similarly, an altered FFA profile, identified by cluster analysis, was related to imbalanced levels of Akkermansia and Lactobacillus as well as increased fecal SCFA, enhanced IL-6 serum levels, and higher prevalence of subclinical metabolic alterations. Although no differences in nutritional intakes were observed, divergent patterns in the associations between nutrient intakes with intestinal microbial populations and SCFA were denoted. Overall, these findings provide new insights on the gut microbiota–host lipid metabolism axis and its potential relevance for human health, where FFA and SCFA seem to play an important role. Keywords: free fatty acids, microbiota, Akkermansia, short-chain fatty acids, serum lipids, subclinical metabolic alterations
INTRODUCTION Free fatty acids (FFA) are lipid species released from the adipose tissue and several cell types upon lipolysis. Apart from their classical roles in energy supply or as structural components, FFA are emerging as active players of a number of biological processes. FFA can affect gene expression of macrophages (1), adipocytes (2), or endothelial cells (3). In addition, FFA can modulate the production of chemokines and cytokines (3–5), the expression of genes coding for adhesion molecules (6, 7) and they give rise to pro-inflammatory and inflammation pro-resolving lipid-derived species (8).
Frontiers in Immunology | www.frontiersin.org
1
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Therefore, a growing body of evidence emphasizes a role for FFA as common mediators between metabolic conditions and the immune system. FFA have been proposed as a mechanistic explanation for the relationship among obesity, inflammation, altered glucose homeostasis, and cardiovascular disease (9). Similarly, some FFA have been reported as regulators of systemic metabolism homeostasis in mice (10). Importantly, quantitative and qualitative differences among the effects of individual FFA have been revealed (11). Thus, not only increased levels but also an altered FFA pool composition may be associated with the risk of developing a range of disorders in which the immune system plays a role (12–17). However, the underlying causes of the altered FFA pool composition remain unknown. Compelling evidence from recent years has shed some light on the effects that gut microbiota can exert on the host health. Alterations in the composition of the intestinal microbiota have been related with the development of metabolic disorders both in mice and in human studies (18), whereas some dietary factors are also known to modulate these microbial communities (19). Recent studies indicate that the gut microbiota is also involved in the host energy metabolism by regulating the absorption of nutrients, local production of hormones and immune mediators, fat storage, and gut permeability (20–22). However, to what extent host lipid metabolism is associated with the intestinal microbial populations, and whether the gut microbiota could be related to FFA levels, remains largely unknown. Short-chain fatty acids (SCFA) are produced in the gut by the metabolic activity of the intestinal microbiota as catabolic end-products from the fermentation of undigested dietary components, mainly complex carbohydrates. These compounds are pivotal in the interactions between the host and intestinal microbial populations (23), but their actual links with the metabolism of fatty acids in humans are controversial. In addition, since most studies have been performed with relatively low sample sizes and mainly with a priori established conditions, evidence for their potential relevance in healthy adults is lacking in the literature. However, due to their pivotal role in shaping the host metabolism, it is tempting to speculate that the intestinal microbiota could be related to FFA levels in healthy populations. Based on these lines of evidence, we hypothesize that alterations in the composition and metabolic activity of the gut microbiota may be associated with altered levels of FFA, which can be in turn be related with some mediators of inflammation. With the aim to gain insight into the joint impact of these players in human health, we have studied the relationships among selected microbial populations, fecal SCFA, serum FFA levels, and composition, as well as inflammatory mediators. In this way, the main objective of the present study was to evaluate whether specific gut microbial signatures can be related to altered levels of FFA species in adults from the general population.
del Principado de Asturias) in compliance with the Declaration of Helsinki. All participants were informed and gave a signed informed consent prior their inclusion in the study.
Subjects
A group of 101 individuals was recruited from the general population in the Asturias region, northern Spain. Exclusion criteria were the diagnosis of chronic or immune-mediated diseases, recent infections, cancer, or altered metabolic conditions, as well as the current usage (within the previous 6 months) of immunomodulatory drugs, metabolic agents, probiotics, or antibiotics. Demographical parameters of the analyzed population are summarized in Table 1.
Table 1 | Description of the study population. Study population (n = 101) Demographical parameters Age, mean (range) Gender (F/M) BMI Biochemical parameters (mg/dl) Glucose Total cholesterol HDL cholesterol LDL cholesterol Triglycerides Microbial populations (log cells/g) Akkermansia Bacteroides group Bifidobacterium Clostridium cluster XIVa Lactobacillus group Faecalibacterium
MATERIALS AND METHODS Ethical Approval
Ethical approval for this study was obtained from the Institu tional Review Board (Comité de Ética de Investigación Clínica
Frontiers in Immunology | www.frontiersin.org
49.41 (19–72) 66/35 26.76 ± 4.39 96.49 ± 12.20 208.66 ± 41.64 56.70 ± 12.84 130.41 ± 35.80 107.53 ± 62.78 5.84 ± 1.80 9.15 ± 1.16 8.26 ± 0.91 8.52 ± 1.41 6.14 ± 1.04 8.05 ± 1.29
Main short-chain fatty acid (SCFA) (mM) Acetate Propionate Butyrate Total SCFA
40.46 ± 19.20 15.04 ± 8.34 10.61 ± 7.19 66.42 ± 31.06
Free fatty acids (FFA) (μg/ml) Palmitic (16:0) Stearic (18:0) Palmitoleic (16:1w7) Oleic (18:1w9) Linoleic (18:2w6) γ-Linoleic (18:3w6) AA (20:4w6) Linolenic (18:3w3) Eicosapentaenoic (20:5w3) Docosahexaenoic acid (22:6w3) SFA MUFA w6-PUFA w3-PUFA Total FFA (mM)
28.34 (12.04) 28.35 (9.98) 0.80 (1.20) 27.38 (20.12) 7.90 (7.11) 0.08 (0.08) 2.02 (1.49) 0.18 (0.16) 0.09 (0.16) 1.31 (1.59) 58.28 (19.08) 28.15 (20.48) 10.59 (8.21) 1.67 (1.74) 0.31 (0.17)
Cytokine serum levels (pg/ml) IL-6 MCP-1 IFNγ
442.79 (949.99) 267.30 (127.83) 2.05 (4.70)
Variables are summarized as median (interquartile range), mean ± SD, or n (%), as appropriated, unless otherwise stated.
2
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Subjects were asked to participate in this study and, upon acceptance, an overnight fast blood sample was drawn by venipuncture in tubes without anticoagulant. After blood clotting, serum was collected and stored at −80°C until further analyses. In addition, basic serum biochemical analyses were performed by standardized procedures. Subjects were considered to exhibit a subclinical metabolic alteration if they meet any of the following criteria: fasting glucose levels higher than 100 mg/dl or that of triglycerides over 150 mg/dl. These objectives cut offs were obtained from clinical national guidelines.
docosahexaenoic acid (DHA) and linolenic; 2.3–75 µg/ml for arachidonic (AA) and palmitoleic; 3.9–125 µg/ml for linoleic; and 7.8–250 µg/ml for oleic, palmitic, and stearic. A good linearity was observed in all cases (r2 > 0.994). Heptadecanoic acid, added at a level of 30 µg/ml, was used as internal standard to compensate for possible biases during the sample preparation step.
Quantification of Cytokines
Serum levels of IFNγ, MCP-1, and IL-6 were determined by immunoassays using commercial kits (IFNγ OptEIA kit) from BD Biosciences (NJ, USA) and MCP-1 and IL-6 mini-EDK kits from Peprotech (NJ, USA), following manufacturer’s instructions. Detection limits were 0.58, 8, and 5.8 pg/ml for IFNγ, MCP-1, and IL-6, respectively.
Total FFA Assessment
Total FFA serum levels were quantified by means of a colorimetric assay using a commercial kit (NEFA kit half-microtest, Roche Life Sciences, Penzberg, Germany) following the protocol from the manufacturer. Final absorbance was measured at 546 nm, and detection limit was 0.02 mM.
Nutritional Assessments
Dietary intake was assessed by means of an annual semi quantitative validated food frequency questionnaire including 160 items (25). Trained dieticians asked about cooking practices, number and amount of ingredients used in each recipe, as well as enquiring about menu preparation (e.g., type of oil used and type of milk) and other relevant information to the study. During an interview, subjects were asked item by item whether they usually ate each food and, if so, how much they usually ate. For this purpose, three different serving sizes of each cooked food were presented in pictures to the participants so that they could choose from up to seven serving sizes (from “less than the small one” to “more than the large one”). For some of the foods consumed, amounts were recorded in household units, by volume, or by measuring with a ruler. Methodological issues concerning dietary assessment have been detailed elsewhere (25). Food intake was analyzed for energy, macronutrients, and total fiber content by using the nutrient Food Composition Tables developed by the Centro de Enseñanza Superior de Nutrición Humana y Dietética (26).
Individual FFA Quantification
Free fatty acids were analyzed after a methyl-tert-butylether (MTBE)-based extraction protocol as previously described (24) with slight modifications. Briefly, 100 µl serum samples were spiked with 5 µl of internal standard (600 ppm heptadecanoic acid). Then, protein precipitation was performed by adding 200 µl methanol chromasolv grade (Sigma Aldrich, MO, USA), and tubes were vortexed for 30 s. Then, 1,200 µl MTBE chromasolv grade (Sigma) was added, vortexed, and an incubation in an ultrasound water bath at 15°C was performed for 30 min. Finally, organic phase was separated after the addition of 200 µl milliQ water and a centrifugation step for 7 min at 5,000 rpm (15°C). After collecting the upper layer, the extraction protocol was repeated once with 100 µl MetOH, 500 µl MTBE, and 100 µl milliQ H2O. Lipid extracts were dried in a miVac centrifugal evaporator (Genevac Ltd., UK) and redissolved in 100 µl of water:acetonitrile 38:62. For the determination of the fatty acids, a Dionex Ultimate 3000 HPLC system (Thermo Scientific, Bremen, Germany), consisting of a high pressure binary pump, an autosampler and a column oven, was used. The column was a Zorbax Eclipse Plus C18, 50 mm × 2.1 mm, 1.8 μm from Agilent. Mobile phases A and B were water and acetonitrile, respectively, both containing 0.1% of formic acid. Fatty acids separation was carried out by the following gradient program: 62% B (held for 4.5 min) followed by a linear increase up to 100% B in 10 min (held for 1 min). The column temperature was set at 45°C and the injection volume was 2 µl. Mass detection was performed using a Bruker Impact II q-ToF mass spectrometer with electrospray ionization, operated in the negative mode. The settings of the mass spectrometer were as follows: spray voltage, 4.5 kV; drying gas flow 12 l/min; drying gas temperature 250°C; nebulizer pressure 44°psi. For quantitation, calibration curves for each compound were prepared by proper dissolution of the pure standards in methanol to encompass the expected concentration of the analytes in the sample. The calibration ranges were as follows: 0.4–12.5 µg/ml for eicosapentaenoic (EPA) and γ-linolenic; 1.2–37.5 µg/ml for Frontiers in Immunology | www.frontiersin.org
Anthropometric Measures
Height was measured using a stadiometer with an accuracy of ±1 mm (Año-Sayol, Barcelona, Spain). The subjects stood barefoot, in an upright position and with the head positioned in the Frankfort horizontal plane. Weight was measured on a scale with an accuracy of ±100 g (Seca, Hamburg, Germany). Quetelet index was calculated using the formula: weight (kg)/height (m2).
Analysis of Fecal Microbiota
Fecal samples from individuals participating in the study were collected at home in sterile containers, kept at 4°C in the domestic refrigerator (maximum 2–3 h from deposition) and frozen at −80°C until analyses just right on arrival to the laboratory, as previously reported (27, 28). One gram of fecal sample was employed for DNA extraction by using the QIAamp DNA stool mini kit (Qiagen, Hilden, Germany) as previously described (28). Quantification of bacterial groups (Akkermansia, Bacteroides group, Bifidobacterium, Clostridium cluster XIVa, Lactobacillus group, and Faecalibacterium) in feces was performed with a 7500 Fast Real Time PCR System (Applied Biosystems, Foster
3
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
City, CA, USA) using SYBR Green PCR Master Mix (Applied Biosystems) as previously described (29–31). These microbial groups include the main representatives of the human intestinal microbiota, which all together represent over 95% of the total bacteria in the human intestine (32, 33). For qPCR analysis, 1 µl of template fecal DNA (~5 ng) and 0.2 µM of each primer were added to the reaction mixture (25 µl). PCR cycling was as follows: an initial cycle of 95°C 10 min, 40 cycles of 95°C 15 s, and 1 min at the appropriate primer temperature. We compared the Ct values obtained from a standard curve constructed as previously indicated (30) to estimate the number of cells. Standard cultures, primers, and annealing temperatures used for qPCR were the same as those recently reported (31).
assess statistical differences, when appropriate. Correlations were analyzed by Spearman’s rank test. After univariate correlation analyses, multiple regression analyses were conducted to assess the strength of the association between these variables including other (continuous) variables as potential confounders. Multiple regression analyses were also performed to identify the main predictors of a candidate independent variable. When coefficients of determination correspond to multiple regression analyses they are indicated as R2. For each analysis, dependent and independent variables were indicated and the β coefficient, B coefficient with 95% confidence intervals (CIs), and p-values were computed. The association between two categorical variables was first studied by χ2 tests in univariate models and then was analyzed by multiple logistic regression models to include potential confounders as independent variables. For this analysis, odds ratio (OR) and 95% CI were computed. Principal component analysis (PCA) with Varimax rotation was performed to reduce sample dimensionality and potential collinearity effects. The number of components retained was based on eigenvalues (>1), and loadings greater than 0.5 were used to identify the variables comprising a single component. For cluster analysis, squared Euclidean distances were calculated, and Ward’s Minimum Variance Method was used to identify clusters minimizing the loss of information. R package heatmap.2 was used to generate heatmaps for visualization purposes. Some statistical analyses were performed independently in each cluster to evaluate whether differences among the studied variables independently arise on each cluster. The statistical approach is summarized in Figure 1. SPSS 19.0, R 3.0.3, and GraphPad Prism 5.0 for Windows were used.
Analysis of Short-Chain Fatty Acids in Fecal Samples
Analysis of short chain fatty acids (SCFA) was performed by gas chromatography to determine the concentrations of acetate, propionate, and butyrate. One gram of fecal samples was weighed, diluted 1:10 in sterile PBS, and homogenized in a LabBlender 400 stomacher (Seward Medical, London, UK) at full speed for 4 min. Supernatants were then obtained by centrifugation (10,000 g, 30 min, 4°C), filtered through 0.2-µm filters, mixed with 1/10 of ethyl butyric (2 mg/ml) as an internal standard, and stored at −80°C until analysis. A gas chromatograph 6890N (Agilent Technologies Inc., Palo Alto, CA, USA) connected to a mass spectrometry (MS) 5973N detector (Agilent Technologies) and to a flame ionization detector was used for identification and quantification of SCFA. Data were collected using the Enhanced ChemStation G1701DA software (Agilent). Samples (1 µl) were directly injected into the gas chromatograph equipped with an HP-Innowax capillary column (60-m length by 0.25-mm internal diameter, with a 0.25 µm film thickness; Agilent) using He as gas carrier and a constant flow rate of 1.5 ml/min. The temperature of the injector was kept at 220°C, and the split ratio was 50:1. Chromatographic conditions were as follows: initial oven temperature of 120°C, 5°C/min up to 180°C, 1 min at 180°C, and a ramp of 20°C/min up to 220°C to clean the column. In the MS detector, the electron impact energy was set at 70 eV. The data collected were in the range of 25–250 atomic mass units (at 3.25 scans/s). SCFA were identified by comparison of their mass spectra with those held in the HP-Wiley 138 library (Agilent) and by comparison of their retention times with those of the corresponding standards (Sigma Aldrich, St. Louis, MO, USA). The peaks were quantified as relative abundances with respect to the internal standard. The concentration (in mM) of each SCFA was calculated using the linear regression equations (r2 ≥ 0.99) from the corresponding standard curves obtained with six different concentrations.
RESULTS FFA and Microbial Populations
Total FFA serum levels and fecal microbial groups were analyzed in 101 human adults from the general population (Table 1). High total FFA serum levels were considered as a surrogate marker of impaired lipid metabolism. Then, the associations between FFA levels and intestinal microbial populations were determined by univariate and multivariate analyses adjusted for socio-demographical [age, gender, and body mass index (BMI)] and nutritional parameters (total energy and intakes of carbohydrates, lipids, proteins, and fiber). Interestingly, only Akkermansia abundance was correlated with FFA serum levels (r = -0.383, p 150 mg/dl) states than those within cluster I (22/39 vs 22/62, p = 0.039 and 7/39 vs 4/62, p = 0.070). Consequently, metabolic alterations (either hyperglycemia or hypertriglycemia, as stated in the Methods section) were more prevalent in cluster II (23/39 vs 24/62, p = 0.047). The strength of these associations was further analyzed by multiple logistic regression. By selecting subclinical metabolic
6
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Figure 3 | Cluster analysis of free fatty acids (FFA) based on specific FFA serum levels. (A) Heatmap showing dendrogram classification for the clusters based on specific FFA serum levels (columns). Colors in the vertical bar at the right of the heatmap identify clusters I (black) and II (gray). Tiles are colored based on serum FFA concentrations, red and blue indicating low or high levels, respectively. (B) Biplot obtained in the cluster analysis of specific FFA serum levels. Red arrows represent the vectors showing the associations among the original variables entered and the outcome of the cluster analysis. Numbers represent the final cluster assigned to each subject (1: cluster I, 2: cluster II). Whereas saturated and mainly pro-inflammatory FA were closely related, EPA and DHA were not involved in clusters definition. (C) Comparison of serum levels of FFA, IL-6, IFNγ, and MCP-1 between subjects of both clusters. Boxes represent median and interquartile range, whereas whiskers represent minimum and maximum values. Differences were assessed by Mann–Whitney U tests.
Table 3 | Demographical parameters according to free fatty acids (FFA) clusters. Study population (n = 101) Demographical parameters Age, mean 49.41 (19–72) (range) Gender (F/M) 66/35 BMI 26.76 ± 4.39
Cluster I (n = 62)
Cluster II (n = 39)
p-Value
50.39 (25–72) 47.78 (19–67)
0.260
42/20 26.67 ± 4.50
0.524 0.580
24/15 26.91 ± 4.27
alterations as the dependent variable and adjusting the model for confounders (age, gender, BMI, total energy, as well as lipids, carbohydrates, protein, and fiber intakes, all introduced as independent variables), we found that individuals within cluster II were more likely to exhibit metabolic alterations (OR [95% CI], p: 3.064 [1.127, 8.330], 0.028). Interestingly, this association remained significant when the microbial groups were entered in the model, appearing Akkermansia abundance also associated with metabolic alterations (p = 0.029). This highlights the interplay between altered FFA profile, microbial populations and impaired host metabolism. Moreover, additional observations reinforced the finding commented just above. Since our initial results point to a cross
Demographical parameters from the whole population (n = 101) and from the study participants classified according to clusters computed from individual FFA levels are shown. Differences between clusters were analyzed by Mann–Whitney U or χ2 tests. Variables are expressed as mean ± SD, median (interquartile range) or n, as appropriate.
Frontiers in Immunology | www.frontiersin.org
7
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Table 4 | Nutritional parameters according to free fatty acids (FFA) clusters. Study population (n = 101)
Cluster I (n = 62)
Cluster II (n = 39)
p-Value (unadjusted)
p-Value (adjusted)
1,993.24 ± 531.62 80.49 ± 27.34 26.24 ± 10.00 33.83 ± 13.25 13.94 ± 9.06 92.15 ± 27.66 213.06 ± 69.43 22.71 ± 10.65 2.75 ± 1.21 14.13 ± 5.88
1,961.40 ± 536.41 80.64 ± 26.31 26.82 ± 10.82 35.08 ± 12.82 12.41 ± 6.24 93.36 ± 28.79 207.73 ± 66.70 23.46 ± 11.74 2.79 ± 1.16 14.50 ± 6.11
2,011.51 ± 515.05 78.23 ± 27.77 25.04 ± 8.54 30.68 ± 11.86 15.91 ± 11.96 88.76 ± 25.58 226.79 ± 73.77 21.75 ± 8.88 2.74 ± 1.32 13.78 ± 5.57
0.660 0.494 0.512 0.100 0.410 0.754 0.125 0.532 0.551 0.523
0.842 0.126 0.109 0.079 0.138 0.132 0.070 0.249 0.972 0.545
Daily intakes Energy (kcal) Lipids (g) Saturated (g) Monounsaturated (g) Polyunsaturated (g) Proteins (g) Carbohydrates (g) Fiber (g) Soluble (g) Insoluble (g)
Dietary intakes from the whole population (n = 101) and from the study participants classified according to clusters computed from individual FFA levels are shown. Differences in daily intakes between clusters were analyzed by Mann–Whitney U tests (unadjusted) or linear regression models to adjust for confounders (adjusted). Energy was adjusted by gender and age, whereas the rest of the nutrients were adjusted by gender, age, and energy. Variables are expressed as mean ± SD.
both cluster I (r = 0.524, p = 0.018) and II (r = 0.432, p = 0.045), but not in those individuals without subclinical metabolic alterations. In sum, these results support an association between microbial populations, nutritional factors, and impaired host metabolism. Subjects exhibiting an altered FFA profile, which also display imbalanced Akkermansia and Lactobacillus populations, were more frequently associated with subclinical pathogenic metabolic states. Divergent patterns in the associations between dietary intakes with microbial populations and SCFA were also observed.
Table 5 | Intestinal microbial populations, fecal short-chain fatty acid (SCFA), and serum cytokine levels according to free fatty acids (FFA) clusters. Study population (n = 101) Microbial populations (log cells/g) Akkermansia 5.84 ± 1.80 Bacteroides group 9.15 ± 1.16 Bifidobacterium 8.26 ± 0.91 Clostridium cluster 8.52 ± 1.41 XIVa Lactobacillus 6.14 ± 1.04 group Faecalibacterium 8.05 ± 1.29 SCFA (mM) Acetate Propionate Butyrate Total SCFA
40.46 ± 19.20 15.04 ± 8.34 10.61 ± 7.19 66.42 ± 31.06
Cytokine levels (pg/ml) IL-6 442.79 (949.99) MCP-1 267.30 (127.83) IFNγ 2.05 (4.70)
Cluster I (n = 62)
Cluster II (n = 39)
p-Value
6.24 ± 1.84 9.05 ± 1.21 8.02 ± 1.44 8.30 ± 1.52
5.09 ± 1.47 9.19 ± 1.17 8.35 ± 0.95 8.63 ± 1.37
0.002 0.587 0.233 0.306
5.92 ± 1.14
6.41 ± 0.86
0.024
7.84 ± 1.15
8.24 ± 1.11
0.105
35.43 ± 14.32 13.03 ± 6.01 9.74 ± 7.07 58.21 ± 23.75
49.31 ± 22.98 18.26 ± 10.45 12.08 ± 7.31 79.66 ± 36.95
0.001 0.003 0.135 0.004
411.72 (636.85) 268.48 (101.13) 2.05 (4.64)
752.40 (1,155.22) 263.20 (128.51) 2.48 (4.12)
0.032
Microbial Imbalanced Populations As Predictors of Impaired FFA Levels
Finally, although Akkermansia abundance was identified as the only microbial predictor of serum FFA levels in the whole population analyzed (n = 101), striking differences in Akkermansia levels were evidenced between clusters I and II. Since the results commented just above suggest that altered microbial populations and their interactions with nutrient intakes may underlie impairment in the host metabolism and that SCFA can have relevance in such interactions, we included all these parameters (microbial populations, SCFA, dietary intakes, age, gender, and BMI) as independent variables in a multivariate regression analysis, FFA levels being selected as the dependent variable, after stratifying our populations by the clusters obtained in the previous analysis. Interestingly, whereas only age (p = 0.038) and BMI (p = 0.050) were identified as predictors of FFA levels in individuals within cluster I (R2 (model) = 0.407), Lactobacillus abundance (p = 0.032), fiber intake (p = 0.021), SCFA (acetate p = 0.008, propionate p = 0.018, and butyrate p = 0.003), and gender (p = 0.008) were associated with FFA levels in subjects within cluster II (R2 (model) = 0.646). In summary, different factors were found to be associated with total FFA levels depending on the levels of Akkermansia. Above a threshold of Akkermansia levels, anthropometric factors such as BMI and age were the main predictors of serum FFA. However, when Akkermansia abundance is low, factors other than the anthropometric characteristics (as imbalanced microbial populations or SCFA production) seem to impact the FFA pool.
0.541 0.252
Microbial populations and SCFA fecal levels and serum cytokines from the whole population (n = 101) and from the study participants classified according to clusters computed from individual FFA levels are shown. Differences in microbial populations, fecal SCFA, and serum cytokine levels between clusters were analyzed by t or Mann– Whitney U tests. The p-values highlighted in bold represent statistically significant differences. Variables are expressed as mean ± SD or median (interquartile range).
talk between host and microbial metabolism, we further deepen into this idea focusing on possible differential associations between microbial groups or SCFA with nutrient intakes. In subjects with subclinical metabolic alterations, Lactobacillus abundance was correlated with carbohydrates intake in both cluster I (r = 0.603, p = 0.004) and II (r = 0.765, p
Free Fatty acids Profiles are related to gut Microbiota signatures and short-chain Fatty acids Javier Rodríguez-Carrio1, Nuria Salazar1, Abelardo Margolles1, Sonia González2, Miguel Gueimonde1, Clara G. de los Reyes-Gavilán1* and Ana Suárez3 Department of Microbiology and Biochemistry of Dairy Products, Instituto de Productos Lácteos de Asturias (IPLA-CSIC), Villaviciosa, Asturias, Spain, 2 Area of Physiology, Department of Functional Biology, University of Oviedo, Oviedo, Asturias, Spain, 3 Area of Immunology, Department of Functional Biology, University of Oviedo, Oviedo, Asturias, Spain 1
Edited by: Marina I. Arleevskaya, Kazan State Medical Academy, Russia Reviewed by: Mario M. D’Elios, University of Florence, Italy Francois-Pierre Martin, Nestle Institute of Health Sciences, Switzerland Arne Yndestad, Oslo University Hospital, Norway *Correspondence: Clara G. de los Reyes-Gavilán [email protected] Specialty section: This article was submitted to Microbial Immunology, a section of the journal Frontiers in Immunology Received: 07 April 2017 Accepted: 29 June 2017 Published: 24 July 2017 Citation: Rodríguez-Carrio J, Salazar N, Margolles A, González S, Gueimonde M, de los ReyesGavilán CG and Suárez A (2017) Free Fatty Acids Profiles Are Related to Gut Microbiota Signatures and Short-Chain Fatty Acids. Front. Immunol. 8:823. doi: 10.3389/fimmu.2017.00823
A growing body of evidence highlights the relevance of free fatty acids (FFA) for human health, and their role in the cross talk between the metabolic status and immune system. Altered serum FFA profiles are related to several metabolic conditions, although the underlying mechanisms remain unclear. Recent studies have highlighted the link between gut microbiota and host metabolism. However, although most of the studies have focused on different clinical conditions, evidence on the role of these mediators in healthy populations is lacking. Therefore, we have addressed the analysis of the relationship among gut microbial populations, short-chain fatty acid (SCFA) production, FFA levels, and immune mediators (IFNγ, IL-6, and MCP-1) in 101 human adults from the general Spanish population. Levels of selected microbial groups, representing the major phylogenetic types present in the human intestinal microbiota, were determined by quantitative PCR. Our results showed that the intestinal abundance of Akkermansia was the main predictor of total FFA serum levels, displaying a negative association with total FFA and the pro-inflammatory cytokine IL-6. Similarly, an altered FFA profile, identified by cluster analysis, was related to imbalanced levels of Akkermansia and Lactobacillus as well as increased fecal SCFA, enhanced IL-6 serum levels, and higher prevalence of subclinical metabolic alterations. Although no differences in nutritional intakes were observed, divergent patterns in the associations between nutrient intakes with intestinal microbial populations and SCFA were denoted. Overall, these findings provide new insights on the gut microbiota–host lipid metabolism axis and its potential relevance for human health, where FFA and SCFA seem to play an important role. Keywords: free fatty acids, microbiota, Akkermansia, short-chain fatty acids, serum lipids, subclinical metabolic alterations
INTRODUCTION Free fatty acids (FFA) are lipid species released from the adipose tissue and several cell types upon lipolysis. Apart from their classical roles in energy supply or as structural components, FFA are emerging as active players of a number of biological processes. FFA can affect gene expression of macrophages (1), adipocytes (2), or endothelial cells (3). In addition, FFA can modulate the production of chemokines and cytokines (3–5), the expression of genes coding for adhesion molecules (6, 7) and they give rise to pro-inflammatory and inflammation pro-resolving lipid-derived species (8).
Frontiers in Immunology | www.frontiersin.org
1
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Therefore, a growing body of evidence emphasizes a role for FFA as common mediators between metabolic conditions and the immune system. FFA have been proposed as a mechanistic explanation for the relationship among obesity, inflammation, altered glucose homeostasis, and cardiovascular disease (9). Similarly, some FFA have been reported as regulators of systemic metabolism homeostasis in mice (10). Importantly, quantitative and qualitative differences among the effects of individual FFA have been revealed (11). Thus, not only increased levels but also an altered FFA pool composition may be associated with the risk of developing a range of disorders in which the immune system plays a role (12–17). However, the underlying causes of the altered FFA pool composition remain unknown. Compelling evidence from recent years has shed some light on the effects that gut microbiota can exert on the host health. Alterations in the composition of the intestinal microbiota have been related with the development of metabolic disorders both in mice and in human studies (18), whereas some dietary factors are also known to modulate these microbial communities (19). Recent studies indicate that the gut microbiota is also involved in the host energy metabolism by regulating the absorption of nutrients, local production of hormones and immune mediators, fat storage, and gut permeability (20–22). However, to what extent host lipid metabolism is associated with the intestinal microbial populations, and whether the gut microbiota could be related to FFA levels, remains largely unknown. Short-chain fatty acids (SCFA) are produced in the gut by the metabolic activity of the intestinal microbiota as catabolic end-products from the fermentation of undigested dietary components, mainly complex carbohydrates. These compounds are pivotal in the interactions between the host and intestinal microbial populations (23), but their actual links with the metabolism of fatty acids in humans are controversial. In addition, since most studies have been performed with relatively low sample sizes and mainly with a priori established conditions, evidence for their potential relevance in healthy adults is lacking in the literature. However, due to their pivotal role in shaping the host metabolism, it is tempting to speculate that the intestinal microbiota could be related to FFA levels in healthy populations. Based on these lines of evidence, we hypothesize that alterations in the composition and metabolic activity of the gut microbiota may be associated with altered levels of FFA, which can be in turn be related with some mediators of inflammation. With the aim to gain insight into the joint impact of these players in human health, we have studied the relationships among selected microbial populations, fecal SCFA, serum FFA levels, and composition, as well as inflammatory mediators. In this way, the main objective of the present study was to evaluate whether specific gut microbial signatures can be related to altered levels of FFA species in adults from the general population.
del Principado de Asturias) in compliance with the Declaration of Helsinki. All participants were informed and gave a signed informed consent prior their inclusion in the study.
Subjects
A group of 101 individuals was recruited from the general population in the Asturias region, northern Spain. Exclusion criteria were the diagnosis of chronic or immune-mediated diseases, recent infections, cancer, or altered metabolic conditions, as well as the current usage (within the previous 6 months) of immunomodulatory drugs, metabolic agents, probiotics, or antibiotics. Demographical parameters of the analyzed population are summarized in Table 1.
Table 1 | Description of the study population. Study population (n = 101) Demographical parameters Age, mean (range) Gender (F/M) BMI Biochemical parameters (mg/dl) Glucose Total cholesterol HDL cholesterol LDL cholesterol Triglycerides Microbial populations (log cells/g) Akkermansia Bacteroides group Bifidobacterium Clostridium cluster XIVa Lactobacillus group Faecalibacterium
MATERIALS AND METHODS Ethical Approval
Ethical approval for this study was obtained from the Institu tional Review Board (Comité de Ética de Investigación Clínica
Frontiers in Immunology | www.frontiersin.org
49.41 (19–72) 66/35 26.76 ± 4.39 96.49 ± 12.20 208.66 ± 41.64 56.70 ± 12.84 130.41 ± 35.80 107.53 ± 62.78 5.84 ± 1.80 9.15 ± 1.16 8.26 ± 0.91 8.52 ± 1.41 6.14 ± 1.04 8.05 ± 1.29
Main short-chain fatty acid (SCFA) (mM) Acetate Propionate Butyrate Total SCFA
40.46 ± 19.20 15.04 ± 8.34 10.61 ± 7.19 66.42 ± 31.06
Free fatty acids (FFA) (μg/ml) Palmitic (16:0) Stearic (18:0) Palmitoleic (16:1w7) Oleic (18:1w9) Linoleic (18:2w6) γ-Linoleic (18:3w6) AA (20:4w6) Linolenic (18:3w3) Eicosapentaenoic (20:5w3) Docosahexaenoic acid (22:6w3) SFA MUFA w6-PUFA w3-PUFA Total FFA (mM)
28.34 (12.04) 28.35 (9.98) 0.80 (1.20) 27.38 (20.12) 7.90 (7.11) 0.08 (0.08) 2.02 (1.49) 0.18 (0.16) 0.09 (0.16) 1.31 (1.59) 58.28 (19.08) 28.15 (20.48) 10.59 (8.21) 1.67 (1.74) 0.31 (0.17)
Cytokine serum levels (pg/ml) IL-6 MCP-1 IFNγ
442.79 (949.99) 267.30 (127.83) 2.05 (4.70)
Variables are summarized as median (interquartile range), mean ± SD, or n (%), as appropriated, unless otherwise stated.
2
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Subjects were asked to participate in this study and, upon acceptance, an overnight fast blood sample was drawn by venipuncture in tubes without anticoagulant. After blood clotting, serum was collected and stored at −80°C until further analyses. In addition, basic serum biochemical analyses were performed by standardized procedures. Subjects were considered to exhibit a subclinical metabolic alteration if they meet any of the following criteria: fasting glucose levels higher than 100 mg/dl or that of triglycerides over 150 mg/dl. These objectives cut offs were obtained from clinical national guidelines.
docosahexaenoic acid (DHA) and linolenic; 2.3–75 µg/ml for arachidonic (AA) and palmitoleic; 3.9–125 µg/ml for linoleic; and 7.8–250 µg/ml for oleic, palmitic, and stearic. A good linearity was observed in all cases (r2 > 0.994). Heptadecanoic acid, added at a level of 30 µg/ml, was used as internal standard to compensate for possible biases during the sample preparation step.
Quantification of Cytokines
Serum levels of IFNγ, MCP-1, and IL-6 were determined by immunoassays using commercial kits (IFNγ OptEIA kit) from BD Biosciences (NJ, USA) and MCP-1 and IL-6 mini-EDK kits from Peprotech (NJ, USA), following manufacturer’s instructions. Detection limits were 0.58, 8, and 5.8 pg/ml for IFNγ, MCP-1, and IL-6, respectively.
Total FFA Assessment
Total FFA serum levels were quantified by means of a colorimetric assay using a commercial kit (NEFA kit half-microtest, Roche Life Sciences, Penzberg, Germany) following the protocol from the manufacturer. Final absorbance was measured at 546 nm, and detection limit was 0.02 mM.
Nutritional Assessments
Dietary intake was assessed by means of an annual semi quantitative validated food frequency questionnaire including 160 items (25). Trained dieticians asked about cooking practices, number and amount of ingredients used in each recipe, as well as enquiring about menu preparation (e.g., type of oil used and type of milk) and other relevant information to the study. During an interview, subjects were asked item by item whether they usually ate each food and, if so, how much they usually ate. For this purpose, three different serving sizes of each cooked food were presented in pictures to the participants so that they could choose from up to seven serving sizes (from “less than the small one” to “more than the large one”). For some of the foods consumed, amounts were recorded in household units, by volume, or by measuring with a ruler. Methodological issues concerning dietary assessment have been detailed elsewhere (25). Food intake was analyzed for energy, macronutrients, and total fiber content by using the nutrient Food Composition Tables developed by the Centro de Enseñanza Superior de Nutrición Humana y Dietética (26).
Individual FFA Quantification
Free fatty acids were analyzed after a methyl-tert-butylether (MTBE)-based extraction protocol as previously described (24) with slight modifications. Briefly, 100 µl serum samples were spiked with 5 µl of internal standard (600 ppm heptadecanoic acid). Then, protein precipitation was performed by adding 200 µl methanol chromasolv grade (Sigma Aldrich, MO, USA), and tubes were vortexed for 30 s. Then, 1,200 µl MTBE chromasolv grade (Sigma) was added, vortexed, and an incubation in an ultrasound water bath at 15°C was performed for 30 min. Finally, organic phase was separated after the addition of 200 µl milliQ water and a centrifugation step for 7 min at 5,000 rpm (15°C). After collecting the upper layer, the extraction protocol was repeated once with 100 µl MetOH, 500 µl MTBE, and 100 µl milliQ H2O. Lipid extracts were dried in a miVac centrifugal evaporator (Genevac Ltd., UK) and redissolved in 100 µl of water:acetonitrile 38:62. For the determination of the fatty acids, a Dionex Ultimate 3000 HPLC system (Thermo Scientific, Bremen, Germany), consisting of a high pressure binary pump, an autosampler and a column oven, was used. The column was a Zorbax Eclipse Plus C18, 50 mm × 2.1 mm, 1.8 μm from Agilent. Mobile phases A and B were water and acetonitrile, respectively, both containing 0.1% of formic acid. Fatty acids separation was carried out by the following gradient program: 62% B (held for 4.5 min) followed by a linear increase up to 100% B in 10 min (held for 1 min). The column temperature was set at 45°C and the injection volume was 2 µl. Mass detection was performed using a Bruker Impact II q-ToF mass spectrometer with electrospray ionization, operated in the negative mode. The settings of the mass spectrometer were as follows: spray voltage, 4.5 kV; drying gas flow 12 l/min; drying gas temperature 250°C; nebulizer pressure 44°psi. For quantitation, calibration curves for each compound were prepared by proper dissolution of the pure standards in methanol to encompass the expected concentration of the analytes in the sample. The calibration ranges were as follows: 0.4–12.5 µg/ml for eicosapentaenoic (EPA) and γ-linolenic; 1.2–37.5 µg/ml for Frontiers in Immunology | www.frontiersin.org
Anthropometric Measures
Height was measured using a stadiometer with an accuracy of ±1 mm (Año-Sayol, Barcelona, Spain). The subjects stood barefoot, in an upright position and with the head positioned in the Frankfort horizontal plane. Weight was measured on a scale with an accuracy of ±100 g (Seca, Hamburg, Germany). Quetelet index was calculated using the formula: weight (kg)/height (m2).
Analysis of Fecal Microbiota
Fecal samples from individuals participating in the study were collected at home in sterile containers, kept at 4°C in the domestic refrigerator (maximum 2–3 h from deposition) and frozen at −80°C until analyses just right on arrival to the laboratory, as previously reported (27, 28). One gram of fecal sample was employed for DNA extraction by using the QIAamp DNA stool mini kit (Qiagen, Hilden, Germany) as previously described (28). Quantification of bacterial groups (Akkermansia, Bacteroides group, Bifidobacterium, Clostridium cluster XIVa, Lactobacillus group, and Faecalibacterium) in feces was performed with a 7500 Fast Real Time PCR System (Applied Biosystems, Foster
3
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
City, CA, USA) using SYBR Green PCR Master Mix (Applied Biosystems) as previously described (29–31). These microbial groups include the main representatives of the human intestinal microbiota, which all together represent over 95% of the total bacteria in the human intestine (32, 33). For qPCR analysis, 1 µl of template fecal DNA (~5 ng) and 0.2 µM of each primer were added to the reaction mixture (25 µl). PCR cycling was as follows: an initial cycle of 95°C 10 min, 40 cycles of 95°C 15 s, and 1 min at the appropriate primer temperature. We compared the Ct values obtained from a standard curve constructed as previously indicated (30) to estimate the number of cells. Standard cultures, primers, and annealing temperatures used for qPCR were the same as those recently reported (31).
assess statistical differences, when appropriate. Correlations were analyzed by Spearman’s rank test. After univariate correlation analyses, multiple regression analyses were conducted to assess the strength of the association between these variables including other (continuous) variables as potential confounders. Multiple regression analyses were also performed to identify the main predictors of a candidate independent variable. When coefficients of determination correspond to multiple regression analyses they are indicated as R2. For each analysis, dependent and independent variables were indicated and the β coefficient, B coefficient with 95% confidence intervals (CIs), and p-values were computed. The association between two categorical variables was first studied by χ2 tests in univariate models and then was analyzed by multiple logistic regression models to include potential confounders as independent variables. For this analysis, odds ratio (OR) and 95% CI were computed. Principal component analysis (PCA) with Varimax rotation was performed to reduce sample dimensionality and potential collinearity effects. The number of components retained was based on eigenvalues (>1), and loadings greater than 0.5 were used to identify the variables comprising a single component. For cluster analysis, squared Euclidean distances were calculated, and Ward’s Minimum Variance Method was used to identify clusters minimizing the loss of information. R package heatmap.2 was used to generate heatmaps for visualization purposes. Some statistical analyses were performed independently in each cluster to evaluate whether differences among the studied variables independently arise on each cluster. The statistical approach is summarized in Figure 1. SPSS 19.0, R 3.0.3, and GraphPad Prism 5.0 for Windows were used.
Analysis of Short-Chain Fatty Acids in Fecal Samples
Analysis of short chain fatty acids (SCFA) was performed by gas chromatography to determine the concentrations of acetate, propionate, and butyrate. One gram of fecal samples was weighed, diluted 1:10 in sterile PBS, and homogenized in a LabBlender 400 stomacher (Seward Medical, London, UK) at full speed for 4 min. Supernatants were then obtained by centrifugation (10,000 g, 30 min, 4°C), filtered through 0.2-µm filters, mixed with 1/10 of ethyl butyric (2 mg/ml) as an internal standard, and stored at −80°C until analysis. A gas chromatograph 6890N (Agilent Technologies Inc., Palo Alto, CA, USA) connected to a mass spectrometry (MS) 5973N detector (Agilent Technologies) and to a flame ionization detector was used for identification and quantification of SCFA. Data were collected using the Enhanced ChemStation G1701DA software (Agilent). Samples (1 µl) were directly injected into the gas chromatograph equipped with an HP-Innowax capillary column (60-m length by 0.25-mm internal diameter, with a 0.25 µm film thickness; Agilent) using He as gas carrier and a constant flow rate of 1.5 ml/min. The temperature of the injector was kept at 220°C, and the split ratio was 50:1. Chromatographic conditions were as follows: initial oven temperature of 120°C, 5°C/min up to 180°C, 1 min at 180°C, and a ramp of 20°C/min up to 220°C to clean the column. In the MS detector, the electron impact energy was set at 70 eV. The data collected were in the range of 25–250 atomic mass units (at 3.25 scans/s). SCFA were identified by comparison of their mass spectra with those held in the HP-Wiley 138 library (Agilent) and by comparison of their retention times with those of the corresponding standards (Sigma Aldrich, St. Louis, MO, USA). The peaks were quantified as relative abundances with respect to the internal standard. The concentration (in mM) of each SCFA was calculated using the linear regression equations (r2 ≥ 0.99) from the corresponding standard curves obtained with six different concentrations.
RESULTS FFA and Microbial Populations
Total FFA serum levels and fecal microbial groups were analyzed in 101 human adults from the general population (Table 1). High total FFA serum levels were considered as a surrogate marker of impaired lipid metabolism. Then, the associations between FFA levels and intestinal microbial populations were determined by univariate and multivariate analyses adjusted for socio-demographical [age, gender, and body mass index (BMI)] and nutritional parameters (total energy and intakes of carbohydrates, lipids, proteins, and fiber). Interestingly, only Akkermansia abundance was correlated with FFA serum levels (r = -0.383, p 150 mg/dl) states than those within cluster I (22/39 vs 22/62, p = 0.039 and 7/39 vs 4/62, p = 0.070). Consequently, metabolic alterations (either hyperglycemia or hypertriglycemia, as stated in the Methods section) were more prevalent in cluster II (23/39 vs 24/62, p = 0.047). The strength of these associations was further analyzed by multiple logistic regression. By selecting subclinical metabolic
6
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Figure 3 | Cluster analysis of free fatty acids (FFA) based on specific FFA serum levels. (A) Heatmap showing dendrogram classification for the clusters based on specific FFA serum levels (columns). Colors in the vertical bar at the right of the heatmap identify clusters I (black) and II (gray). Tiles are colored based on serum FFA concentrations, red and blue indicating low or high levels, respectively. (B) Biplot obtained in the cluster analysis of specific FFA serum levels. Red arrows represent the vectors showing the associations among the original variables entered and the outcome of the cluster analysis. Numbers represent the final cluster assigned to each subject (1: cluster I, 2: cluster II). Whereas saturated and mainly pro-inflammatory FA were closely related, EPA and DHA were not involved in clusters definition. (C) Comparison of serum levels of FFA, IL-6, IFNγ, and MCP-1 between subjects of both clusters. Boxes represent median and interquartile range, whereas whiskers represent minimum and maximum values. Differences were assessed by Mann–Whitney U tests.
Table 3 | Demographical parameters according to free fatty acids (FFA) clusters. Study population (n = 101) Demographical parameters Age, mean 49.41 (19–72) (range) Gender (F/M) 66/35 BMI 26.76 ± 4.39
Cluster I (n = 62)
Cluster II (n = 39)
p-Value
50.39 (25–72) 47.78 (19–67)
0.260
42/20 26.67 ± 4.50
0.524 0.580
24/15 26.91 ± 4.27
alterations as the dependent variable and adjusting the model for confounders (age, gender, BMI, total energy, as well as lipids, carbohydrates, protein, and fiber intakes, all introduced as independent variables), we found that individuals within cluster II were more likely to exhibit metabolic alterations (OR [95% CI], p: 3.064 [1.127, 8.330], 0.028). Interestingly, this association remained significant when the microbial groups were entered in the model, appearing Akkermansia abundance also associated with metabolic alterations (p = 0.029). This highlights the interplay between altered FFA profile, microbial populations and impaired host metabolism. Moreover, additional observations reinforced the finding commented just above. Since our initial results point to a cross
Demographical parameters from the whole population (n = 101) and from the study participants classified according to clusters computed from individual FFA levels are shown. Differences between clusters were analyzed by Mann–Whitney U or χ2 tests. Variables are expressed as mean ± SD, median (interquartile range) or n, as appropriate.
Frontiers in Immunology | www.frontiersin.org
7
July 2017 | Volume 8 | Article 823
Rodríguez-Carrio et al.
FFA and Gut Microbiota
Table 4 | Nutritional parameters according to free fatty acids (FFA) clusters. Study population (n = 101)
Cluster I (n = 62)
Cluster II (n = 39)
p-Value (unadjusted)
p-Value (adjusted)
1,993.24 ± 531.62 80.49 ± 27.34 26.24 ± 10.00 33.83 ± 13.25 13.94 ± 9.06 92.15 ± 27.66 213.06 ± 69.43 22.71 ± 10.65 2.75 ± 1.21 14.13 ± 5.88
1,961.40 ± 536.41 80.64 ± 26.31 26.82 ± 10.82 35.08 ± 12.82 12.41 ± 6.24 93.36 ± 28.79 207.73 ± 66.70 23.46 ± 11.74 2.79 ± 1.16 14.50 ± 6.11
2,011.51 ± 515.05 78.23 ± 27.77 25.04 ± 8.54 30.68 ± 11.86 15.91 ± 11.96 88.76 ± 25.58 226.79 ± 73.77 21.75 ± 8.88 2.74 ± 1.32 13.78 ± 5.57
0.660 0.494 0.512 0.100 0.410 0.754 0.125 0.532 0.551 0.523
0.842 0.126 0.109 0.079 0.138 0.132 0.070 0.249 0.972 0.545
Daily intakes Energy (kcal) Lipids (g) Saturated (g) Monounsaturated (g) Polyunsaturated (g) Proteins (g) Carbohydrates (g) Fiber (g) Soluble (g) Insoluble (g)
Dietary intakes from the whole population (n = 101) and from the study participants classified according to clusters computed from individual FFA levels are shown. Differences in daily intakes between clusters were analyzed by Mann–Whitney U tests (unadjusted) or linear regression models to adjust for confounders (adjusted). Energy was adjusted by gender and age, whereas the rest of the nutrients were adjusted by gender, age, and energy. Variables are expressed as mean ± SD.
both cluster I (r = 0.524, p = 0.018) and II (r = 0.432, p = 0.045), but not in those individuals without subclinical metabolic alterations. In sum, these results support an association between microbial populations, nutritional factors, and impaired host metabolism. Subjects exhibiting an altered FFA profile, which also display imbalanced Akkermansia and Lactobacillus populations, were more frequently associated with subclinical pathogenic metabolic states. Divergent patterns in the associations between dietary intakes with microbial populations and SCFA were also observed.
Table 5 | Intestinal microbial populations, fecal short-chain fatty acid (SCFA), and serum cytokine levels according to free fatty acids (FFA) clusters. Study population (n = 101) Microbial populations (log cells/g) Akkermansia 5.84 ± 1.80 Bacteroides group 9.15 ± 1.16 Bifidobacterium 8.26 ± 0.91 Clostridium cluster 8.52 ± 1.41 XIVa Lactobacillus 6.14 ± 1.04 group Faecalibacterium 8.05 ± 1.29 SCFA (mM) Acetate Propionate Butyrate Total SCFA
40.46 ± 19.20 15.04 ± 8.34 10.61 ± 7.19 66.42 ± 31.06
Cytokine levels (pg/ml) IL-6 442.79 (949.99) MCP-1 267.30 (127.83) IFNγ 2.05 (4.70)
Cluster I (n = 62)
Cluster II (n = 39)
p-Value
6.24 ± 1.84 9.05 ± 1.21 8.02 ± 1.44 8.30 ± 1.52
5.09 ± 1.47 9.19 ± 1.17 8.35 ± 0.95 8.63 ± 1.37
0.002 0.587 0.233 0.306
5.92 ± 1.14
6.41 ± 0.86
0.024
7.84 ± 1.15
8.24 ± 1.11
0.105
35.43 ± 14.32 13.03 ± 6.01 9.74 ± 7.07 58.21 ± 23.75
49.31 ± 22.98 18.26 ± 10.45 12.08 ± 7.31 79.66 ± 36.95
0.001 0.003 0.135 0.004
411.72 (636.85) 268.48 (101.13) 2.05 (4.64)
752.40 (1,155.22) 263.20 (128.51) 2.48 (4.12)
0.032
Microbial Imbalanced Populations As Predictors of Impaired FFA Levels
Finally, although Akkermansia abundance was identified as the only microbial predictor of serum FFA levels in the whole population analyzed (n = 101), striking differences in Akkermansia levels were evidenced between clusters I and II. Since the results commented just above suggest that altered microbial populations and their interactions with nutrient intakes may underlie impairment in the host metabolism and that SCFA can have relevance in such interactions, we included all these parameters (microbial populations, SCFA, dietary intakes, age, gender, and BMI) as independent variables in a multivariate regression analysis, FFA levels being selected as the dependent variable, after stratifying our populations by the clusters obtained in the previous analysis. Interestingly, whereas only age (p = 0.038) and BMI (p = 0.050) were identified as predictors of FFA levels in individuals within cluster I (R2 (model) = 0.407), Lactobacillus abundance (p = 0.032), fiber intake (p = 0.021), SCFA (acetate p = 0.008, propionate p = 0.018, and butyrate p = 0.003), and gender (p = 0.008) were associated with FFA levels in subjects within cluster II (R2 (model) = 0.646). In summary, different factors were found to be associated with total FFA levels depending on the levels of Akkermansia. Above a threshold of Akkermansia levels, anthropometric factors such as BMI and age were the main predictors of serum FFA. However, when Akkermansia abundance is low, factors other than the anthropometric characteristics (as imbalanced microbial populations or SCFA production) seem to impact the FFA pool.
0.541 0.252
Microbial populations and SCFA fecal levels and serum cytokines from the whole population (n = 101) and from the study participants classified according to clusters computed from individual FFA levels are shown. Differences in microbial populations, fecal SCFA, and serum cytokine levels between clusters were analyzed by t or Mann– Whitney U tests. The p-values highlighted in bold represent statistically significant differences. Variables are expressed as mean ± SD or median (interquartile range).
talk between host and microbial metabolism, we further deepen into this idea focusing on possible differential associations between microbial groups or SCFA with nutrient intakes. In subjects with subclinical metabolic alterations, Lactobacillus abundance was correlated with carbohydrates intake in both cluster I (r = 0.603, p = 0.004) and II (r = 0.765, p
